Introduction

The 1970s was marked by a general pessimism in the belief that offender recidivism could be reduced through rehabilitation. This pessimism was triggered by a review of the treatment literature that concluded that there ‘little reason to hope that we have … found a sure way of reducing recidivism through rehabilitation’ (Martinson, 1974, p 50). Thus was launched the so-called ‘nothing works’ era, with its focus on deterrence and getting tough on offenders. Not everyone accepted the nothing works doctrine and the conclusion by Martinson was soon challenged by narrative literature reviews that concluded treatment to be effective (Gendreau and Ross, 1979; Ross and Gendreau, 1980).

Although various reviews of the treatment literature in the 1970s and 1980s reached the same general conclusion that some treatments actually did reduce recidivism, there was little understanding of the circumstances under which treatment was most effective. Then in 1990 Andrews and colleagues published the outline of what is known as the Risk-Need-Responsivity (RNR) model of offender assessment and rehabilitation (Andrews et al, 1990a). The paper was based on a selective review of the treatment literature that found treatment to be most effective if it followed three principles. The first principle, the risk principle, stated that the intensity of treatment should be proportional to the risk level of the offender, with the higher-risk offender receiving the most treatment and the lower-risk offender less treatment. The need principle distinguished between criminogenic needs and non-criminogenic needs. Criminogenic needs are related to offending (for example, procriminal attitudes, criminal companions) and non-criminogenic needs unrelated to criminal behaviour (for example, poor self-esteem, neurotic anxiety). When criminogenic needs are targeted in treatment, recidivism is reduced. Finally, the responsivity principle argued for delivering treatment in a way that is meaningful to the offender, and in general cognitive-behavioural interventions are the best suited for most offenders.

An empirical test of the RNR model followed shortly after with a meta-analysis of offender rehabilitation programmes that found adherence to all three principles was associated with a 35-percentage point reduction in recidivism when delivered in the community and 17 percentage points when delivered in residential/custodial settings (Andrews et al, 1990b).